« Superalignement » : différence entre les versions
(Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Superalignment ''' <small> [https://opendatascience.com/openai-has... ») |
Aucun résumé des modifications |
||
| Ligne 13 : | Ligne 13 : | ||
[https://opendatascience.com/openai-has-introduced-superaligment-an-attempt-to-solving-important-issues-related-to-superintelligence/ Source : opendatascience ] | [https://opendatascience.com/openai-has-introduced-superaligment-an-attempt-to-solving-important-issues-related-to-superintelligence/ Source : opendatascience ] | ||
OpenAI has unveiled its latest initiative called “Superalignment” which aims to create robust and scalable methods for aligning these systems with human values in anticipation of Superintelligence. For those who might not know, Superintelligence is a highly advanced form of AI. However, the sheer power of superintelligence also carries inherent risks. | |||
[[Catégorie:vocabulary]] | [[Catégorie:vocabulary]] | ||
Version du 24 juillet 2023 à 08:27
en construction
Définition
XXXXXXXXX
Français
XXXXXXXXX
Anglais
Superalignment
OpenAI has unveiled its latest initiative called “Superalignment” which aims to create robust and scalable methods for aligning these systems with human values in anticipation of Superintelligence. For those who might not know, Superintelligence is a highly advanced form of AI. However, the sheer power of superintelligence also carries inherent risks.
Contributeurs: Patrick Drouin, wiki
